humanish 0.45.0 → 0.47.0

package/docs/contracts/schemas.md

@@ -3,7 +3,7 @@
 Date: 2026-06-02 (current-state note updated 2026-07-14)
 
 Status: reference map for the major contracts shipped through source version
-`0.45.0`; it is not an exhaustive inventory of command/result envelopes. Exported types,
+`0.47.0`; it is not an exhaustive inventory of command/result envelopes. Exported types,
 schema constants, parsers, and validators in `src/` are authoritative. Rows
 marked "reserved" name layering intent only — no code emits or validates them
 yet. Do not emit a reserved schema.
@@ -724,7 +724,11 @@ Core-owned fields:
   budget AFTER productive activity — status `passed`, a NON-FAILURE completion,
   distinct from `timed_out`, which stays reserved for a zero-progress deadline
   hit and remains a failure)
-- `ids`, `counts`, `items[]`, optional `tokenUsage`, `capabilities`
+- `ids`, `counts`, `items[]`, optional `tokenUsage`, `capabilities`. Items may
+  carry `at` (ISO-8601 recording stamp from the loop's clock) and, on
+  click-like `ui_action` items, structured `coord` (`x`/`y`) — both additive
+  (#441): absent on older bundles and non-stamping producers, and absence means
+  "timing/position unrecorded", never zero
 - optional `affordanceUse` (`humanish.affordance-use.v1`): which KIND of route this
   actor took (see Affordance Use below)
 - optional `estimatedCost` (`humanish.actor-estimated-cost.v1`): a token-derived

package/docs/goals/current.md

@@ -29,7 +29,7 @@ study completed, reproduced, and produced a real accessibility finding via a
 keyboard-first participant
 ([docs/goals/email-gated-signup/receipts/](email-gated-signup/receipts/)).
 
-## Current Program Truth (source `0.45.0`)
+## Current Program Truth (source `0.47.0`)
 
 The package source and repository implementation in this tree agree on these
 points:
@@ -467,44 +467,44 @@ Observer hero from a commit-pinned public application (drawDB) shipped in
 `0.16.0`. That run treats the application as a study subject, not a Humanish
 adopter, and does not imply endorsement; it does not satisfy the depth gate.
 
-The active engineering proof slice is the first depth-axis deletion branch: run
-an existing first-party bespoke sim through current Humanish, compare the
-decision outputs, and delete its bespoke generic harness while retaining only a
-thin product extension. The current front is an agent-facing terminal-product
-harness, whose thin-extension budget is being surfaced as engine gaps. Landed so
-far (`0.32.0`–`0.34.0`): the honest-scoring fixes (persona traits now shape the
-terminal agent; a quoted-subject-copy blocker false-positive; a comms false-green
-guard), a config-loadable adopter scorer seam (`review.scorer.ref` /
-`--scorer`), and the full-transcript scorer evidence channel (the scoring
-context carries the whole normalized session, not a ~2KB tail — an adopter
-rubric can no longer prove less than the persisted transcript). Remaining
-before the deletion, in order: pin the decision-equivalence comparator (a
-written contract — which decision outputs, compared how, over which study set —
-judged over fixed transcripts, never fresh-runs-vs-historical rate
-comparisons), rescore the reference study's arms over full evidence, then
-terminal-lane fan-out at roster scope (per-lane missions/personas; a written
-roll-up verdict contract that separates harness-verdict from per-lane mission
-verdicts precedes the fan-out code). The product-credential channel's payment
-half is explicitly OFF this deletion's path — the bespoke harness being deleted
-never transacted, so decision-equivalence does not require it; only its
-host-side measured-cost half (costProbe run-context) remains in scope. Then run
-that harness's study through current Humanish, compare the decision outputs,
-and delete its bespoke engine. Converge on one such branch to
-decision-equivalent green before starting the next archetype (a multi-actor
-shared-state harness) or the breadth panel.
-
-After that checkpoint, re-evaluate this order rather than treating it as an
-automatic queue:
-
-1. centralize run identity, history, and provider ownership behind tested
-   `RunStore` and `ResourceLease` interfaces, and make generated first-run
-   source load-bearing or explicitly optional;
-2. complete the agent-skill deletion branch and use it to force the real
-   external extension budget;
-3. implement multi-origin shared-world only if its adopter gate opens and the
-   ratified packet passes maintainer review;
-4. complete the multi-app adopter deletion branch;
-5. establish the external stratified panel.
+The operator-pinned train-of-thought thread (#427, both stages) and the
+#441 queue head are done as of `0.47.0`, live-proven the same day they
+shipped: trace items carry recording stamps and structured click
+coordinates; the live path flushes `liveActor` partials mid-run so the
+attached Observer's timeline grows while a study runs; playback runs at
+the participant's recorded pace when stamps exist (and the transport says
+which clock it is on); the participant card's decide-line ticks the
+newest reported thought during a live lane; and a live-shaped golden cut
+from a kept receipt run pins the whole shape (`?fixture=live`). On the way
+the capture exposed and closed two evidence-quality defects: the persona
+identity leak (#452) and the credible-pass guard's incentive inversion
+against post-success defect reports (#453 — the verdict scan and the
+friction tally are now separate reads of the same narrative). Shared-world
+roll-up honesty (#364) closed via an external contribution that separates
+credibility, mission endpoint, and convergence into three explicit claims.
+
+The standing queue, in rough order:
+
+1. Observer deep links (#441's last item): hash routing for run/lane/frame —
+   the declared parity regression from the renderer cutover and the
+   prerequisite for #428's cite-turns flags;
+2. registry promotions: the wordmark (#431), the participant card, and the two
+   vendored Base UI wrappers (drawer, popover) once a second surface consumes
+   them;
+3. the adoption cluster: vendor-native key resolution (#436) and the optional
+   desktop peer-dep gap (#346) — the two biggest remaining "npx works
+   first-try" violations;
+4. shared-world honesty, remaining half: per-action evidence (#365) and
+   exposure-flag coverage (#446);
+5. the stakeholder TUI (#455): research + token-translated mocks on the
+   operator review surface, design sign-off gated before any code.
+
+The depth-axis deletion front (an adopter's bespoke terminal-product sim,
+comparator contract posted on the adopter's tracker) is paused awaiting the
+operator's decision-ledger sign-off, deliberately: it resumes on that
+sign-off, not by default. The multi-actor shared-state adopter gate (#166)
+needs one live rerun of its replacement-critical family on current humanish;
+the recon (family, runner pin) is recorded and it runs as its own arc.
 
 The version-pinned README hero is the drawDB real-application study: a live
 four-persona capture that proves package/Observer rendering, public-safe asset

package/docs/ramp/README.md

@@ -2,7 +2,7 @@
 
 Status: public-safe contributor and agent ramp.
 
-Package/source version in this tree: `0.45.0` (2026-08-16). The Observer is phone-usable as a stated requirement (observer/AGENTS.md); interactive primitives start from Base UI. The Observer renderer is the observer/ workspace artifact only; the legacy string-concat renderer was deleted at cutover (#426), and rollback is a version pin to 0.42.0. The containment boundary introduced in
+Package/source version in this tree: `0.47.0` (2026-08-17). The Observer is phone-usable as a stated requirement (observer/AGENTS.md); interactive primitives start from Base UI. The Observer renderer is the observer/ workspace artifact only; the legacy string-concat renderer was deleted at cutover (#426), and rollback is a version pin to 0.42.0. The containment boundary introduced in
 `0.15.1` remains in force: managed run and output paths bind to validated
 physical filesystem identities, and stored provider IDs are evidence, not
 cleanup authority. The bundled OSS meta-lab is dry-run only until

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "humanish",
-  "version": "0.45.0",
+  "version": "0.47.0",
   "description": "Open-source-safe CLI for persona simulation, observer review, and public-safe feedback drafts.",
   "author": "Daniel G Wilson <daniel@danielgwilson.com>",
   "keywords": [
@@ -58,7 +58,7 @@
     "release:check": "pnpm check && pnpm public-surface:scan && pnpm skill:check && npm pack --dry-run",
     "skill:check": "DISABLE_TELEMETRY=1 npx skills add . --list",
     "test": "vitest run",
-    "typecheck": "tsc -p tsconfig.json --noEmit",
+    "typecheck": "tsc -p tsconfig.json --noEmit && pnpm --filter humanish-observer typecheck",
     "humanish:doctor": "pnpm humanish -- doctor",
     "humanish:run": "pnpm humanish -- run --dry-run",
     "humanish:watch": "pnpm humanish -- watch",